Now I've measured a bit. This time my Cosmo was at 15% battery level at the start of charging. During charging, the Cosmo had its lid closed and sat on a wooden desktop surface, with the CoDi facing up. The ambient temperature was around 24°C. I used a Satechi type USB-C power meter to track supply voltage and current. At certain times I briefly opened the Cosmo and checked the battery temperature (and charge level) in an old app called AndroSensor. This is what I found:
00min 8.0V 1.90A 15% 30°C
05min 8.0V 1.90A 20% 30°C
20min 8.0V 1.80A 39% 36°C
30min 8.9V 1.50A 52% 37°C
50min 8.9V 1.20A 76% 35°C
60min 8.9V 0.95A 86% 34°C
70min 8.9V 0.80A 93% 33°C
My readings from 10 and 40 minutes were incomplete (I got distracted), but the surrounding readings, seems to suggest nothing of particular interest happened at those times anyway.
The power supply is rated for 1.67A at 9V, but initially, the Cosmo manages to pull nearly 2A from it. I'd guess that's why the supply voltages drops to around 8V. At around 21 minutes / 40% charge, the current dropped from 1.80A to 1.67A while the voltage rose from around 8.0V to 8.9. As both 8V x 1.9A and 9V x 1.67A are close to 15W, that seems to be the most the power supply can deliver. The rather sudden change in voltage and current, makes me wonder if the Cosmo, somewhat incorrectly tries to initially pull 2A at 9V and then, around 40% charge level steps down to 1.67A, at which point the supply stabilizes at (close to) 9V.
The peak temperature I measured was 37°C around half an hour into the charging. As it had dropped to 35°C at 50 minutes, it seems unlikely it kept climbing much, but, say a "true peak" of 38°C around 35 minutes would not be inconsistent with my rather sparse data. As far as I can tell, my Cosmo felt about as toasty at 30 minutes as I've begun noticing occasionally on V20. Still, that's 10°C below what klampfenfreak measured, which makes me wonder, did that Cosmo sit on something somewhat insulating, say, a sofa?
I would perhaps have been better off using some kind of continuous logging of the battery temperature, though I suspect Android would throttle that as soon as I closed the lid. Keeping the lid open and the screen on, would solve that, but then I'd be measuring "charging while in use" rather than "charging in standby". On that note, when I opened the lid to take a reading, the current tended to rise 0.2A or so (once the supply had that kind of margin to offer), so it seems the charging process tries to pull as much power when Android is awake as when it sleeps, so unless the Cosmo manages to vent heat more efficiently when the lid is open (which I suspect it does), charging while in use, might contribute to a higher battery temperature.
I'm a bit curious as to what particular app klampfenfreak used to monitor the battery. If it manages to keep running while the lid is closed, the measurement itself might contribute to higher readings, as the SoC, as expected, gets noticeably warm, even at fairly light, but prolonged loads.
